It has been demonstrated that controlling a neuroprosthesis is possible by selectively conditioned firings of a neuron group consisting of some neurons. It is also postulated that neuron groups are able to be operantly conditioned to modulate their firings in isolation with neighboring neuron groups. However, it remains unclear if transfer of the operantly conditioned firings between neuron groups occurs. It has been said that the most critical problem of neural operant conditioning using a small number of neurons is a limited life as sources of signals for volitional motor outputs in controlling a neuroprosthesis. In this respect, transfer of operantly conditioned volitional firings in larger neuronal populations is profitable to compensate the limited life of source signals. In this study, multiple neuron groups are recorded by 2 bundles of microwires implanted in the motor cortex of behaving rats. Experiment sessions start after firings of at least 2 neuron groups are identified in the rat cortex. In session 1, the rats are trained to perform a behavioral free-operant task in which nose-poke behaviors are rewarded. During session 2, the rats are rewarded whenever firings of the neuron group 1 detected in real-time with a brain-machine interface (BMI) system satisfy preset criteria. After learning the neural operant task in session 2, the rats conduct the identical operant task using firings of the neuron group 2 during session 3. Data of each of the neuron groups are analyzed offline. Some indexes, like reward rates and delays to reward delivery, are employed to examine the hypothesis that neuronal operant learning is transferred from the neuron group 1 to the neuron group 2. A preliminary results examining the hypothesis will be reported. The result of this study is expected to contribute to advances in neurorehabilitation and neuroprosthesis.